Error run LES when restart from RANS

[Ruonan]

Hello Yvan,

Thank you very much for debugging. Unfortunately, this error occurred again. I use the new feature version 7.1.

I found when I restart LES Smagorinsky from k-omega SST, it's OK. But when I restart LES Smagorinsky from Spalart-Allmaras, this error happened again. So there may be something to do with the S-A model.

Attached please find a very small test case, without any user subroutine. You maybe reproduce this error in 1 minute.

Could you please check and maybe debug it? As we know, restarting LES from RANS is quite useful.

Many thanks,
Ruonan

[Yvan Fournier]

Hello,

I won't have time to do this before the end of next week, so it may be useful to remind me then

I suspect non-handled combination similar to what we had with k-omega (we do not use SA much internally, to the best of my knowledge, so nobody may have noticed this on our side).

Best regards,

Yvan

[Ruonan]

Hello Yvan,

Thank you very much! It will be my pleasure to remind you then.

I've been using k-omega SST before. It was good but when I run in a rotating reference frame (add the Coriolis source term in), I found the result was wrong. I assume it needs some revision, a possible example is shown in the link below. That's why I changed to the S-A model.

https://arc.aiaa.org/doi/abs/10.2514/6.1998-2554

Best regards,
Ruonan

[Yvan Fournier]

Hello,

I looked into the restart code, and it seems none of the various turbulence models can be restarted from Spalart-Allmaras.
For the initialization of LES from k-epsilon or k-omega, I need the turbulent viscosity (computed from k and epsilon or omega, directly accessible in SA), but also need an evaluation of the turbulent kinetic energy, and don't know how to reconstruct that from the SA model's single turbulent viscosity. I'll need to check with colleagues with more advanced turbulence modeling knowledge than I do how best to handle this. If you have a suggestion, it is welcome too.

Best regards,

Yvan

[Ruonan]

Hello Yvan,

Thank you very much for your reply!

The problem you mentioned is very important, but I'm not very good at turbulence modelling. I will talk with my supervisors to see if they have any comments on this.

Best regards,
Ruonan